Mining Frequent Temporal Structures in Web Graph

In most of the user applications today, the web has become a fundamental information resource. Mining the graph structural data from a complex source is a difficult process and in particular when the nature of graph is dynamic. For supporting the mining techniques on large real world network, an exquisite way is to assemble it as a sequences of random graphs such that it extracts the structural relationships. In this paper, we explore the frequent temporal substructures using the pattern growth methodology on temporal web graph. We investigate the properties of frequent subgraphs on a set of random graphs generated at different instances within a time interval where some probability is assigned to each graph. The resultant subgraphs indicates that there are frequent temporal substructures in web graph, which helps in many graph processing applications such as summarization of dynamic web graphs, graph classification, graph clustering, graph indexing and graph similarity.